| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
This reduces the risk for some bugs, for instance due to
signed/unsigned mismatches and unused variables.
|
|
|
|
|
| |
This stabilises flow allocation now that the rbuffs are created upon
flow allocation. Only the IRMd can sync this process sufficiently.
|
|
|
|
|
|
|
|
| |
The fast path will now use an incoming ring buffer per flow per
process. This necessitated the development of a new method for the
asynchronous io call, which is now based on an event queue system for
scalability (fqueue). The ipcpd's and tools have been updated to this
API.
|
|
|
|
|
| |
Both the N and N-1 entity must now call flow_dealloc for the port_id
to be released.
|
|
|
|
|
| |
Flow deallocation from the application will immediately return (void
call). The IRMd will not send a reply message.
|
|
|
|
|
|
|
|
|
|
|
|
| |
IPCPs can now use ap_init() to initialize the memory. All flows are
accessed using flow descriptors, this greatly simplifies IPCP
development. Reverts the fast path to a single ap_rbuff per process.
Splits lib/ipcp into irmd/ipcp and lib/ipcp-dev. Adds a lib/shim-dev
holding tailored functions for shims. Moves the buffer_t to utils.h.
Fixes the shim-eth-llc length field. Removes the flow from shared.h.
Fixes #4
Fixes #5
|
|
|
|
|
| |
Permissions are now set correctly upon creation, removing the need to
call fchmod.
|
|
|
|
|
| |
If there are unregistered IPCPs (no DIF name set) the lookup should
skip over them without segfaulting.
|
|
|
|
| |
Fast path is split in north and southbound paths.
|
| |
|
|
|
|
|
| |
Flow will now allocate correctly when the name is registered in a DIF
before it is bound to an AP.
|
|
|
|
|
|
|
|
|
|
| |
The shm_du_map is renamed to shm_rdrbuff to reflect the Random
Deletion Ringbuffer used in the implementation. The close_on_exit call
is removed and SDUs are cleaned up by the application in the ap_fini()
call. This required a non-blocking peek() operation in the shm_ap_rbuff.
Some initial implementation for future support of qos cubes has been
added to the shm_rdrbuff.
|
|
|
|
| |
Prevents double destruction.
|
|
|
|
|
| |
Flows are deallocated first, then the socket is closed to prevent
IPCPs hanging on the connect.
|
| |
|
| |
|
|
|
|
| |
ap_table had no function to look for entries by apn.
|
| |
|
|
|
|
|
| |
The reg_entry_del_api must be called to correctly set the flow
handling state instead of just clearing the entry from the list.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The complete data model inside the IRMd has been restructured. The
bind operation was revised to allow binding of AP and AP instances and
register those names with different DIFs (see "irm bind" for details).
Server applications require to call ap_init with a server name
argument, client application that do not the be reachable over any DIF
can call ap_init(NULL). Calling ap_init for a client with a specified
name will not have adverse consequences for the application, but will
consume unnecessary resources in the IRMd.
Application servers can now be started at any point after the IRMd has
been started. Starting servers, binding AP names and registering names in
DIFs can be performed in any order that does not defy temporal logic.
Supports naming instances by their pid. In case of IPCP Instances
created with the IRM tool, the name assigned during "irm ipcp create"
can be used.
All the changes required updates in the tools.
|
|\
| |
| |
| | |
lib, irmd: Bind AP instances to AP_subsets
|
| |
| |
| |
| |
| |
| |
| | |
This call will allow grouping AP instances of a certain AP together
which are configured identically. Adds the bind operation to dev and
updates the applications to make use of this call. Flow_alloc is now
only called with the pid and doesn't send the apn anymore.
|
|/
|
|
|
|
| |
This will add a timeout to the socket so that a process won't be
blocked by the actions of the process with which it is communicating
over the socket.
|
|\ |
|
| |\
| | |
| | |
| | | |
ipcp, irmd, lib: Notify IRMd upon IPCP initialization
|
| | |
| | |
| | |
| | |
| | |
| | | |
This will notify the IRMd when the IPCP is initialized and ready to
receive messages. Previously a bootstrap could fail since the IPCP was
not listening to the socket yet.
|
| |\ \
| | | |
| | | |
| | | | |
Several small fixes and enhancements
|
| | |/
| | |
| | |
| | |
| | |
| | |
| | | |
If an IPCP crashes (due to a segfault for instance), it is removed
from the spawned apis list. However, if it was an IPCP it should also
be removed from the IPCPs list, since else on shutdown, the irmd will
try to destroy the IPCP that crashed.
|
|/ /
| |
| |
| |
| |
| | |
Threads from the threadpool will now check the state of the IRMd and
exit gracefully without a need to cancel them. This avoids the unsafe
call of pthread_cancel in the signal handler.
|
|/
|
|
| |
registry_sanitize_apis should be called under write locked reg_lock.
|
| |
|
| |
|
|
|
|
|
| |
Fix missing set of api upon flow_alloc.
Various locking fixes.
|
|
|
|
| |
Locks should be kept before calling cond_wait.
|
|
|
|
|
|
|
| |
Refactors the IRMd to extract reg_api and irm_flow structures to their
own sources.
Fixes some locking bugs.
|
|
|
|
| |
The req_ae_name is freed when the unpacked message is cleaned up.
|
| |
|
|
|
|
|
| |
When killing the oping client before it sent an SDU, some bad
operations were being performed.
|
| |
|
|
|
|
|
| |
When a flow was in FLOW_ALLOCATED state, it would not change to NULL
state and irm_flow_destroy would hang forever.
|
|
|
|
| |
ipcp-data was not correctly destroyed.
|
|
|
|
|
|
|
|
|
|
| |
Various portability fixes for FreeBSD. POSIX requires shm file names
to start with a "/" to be portable. lseek(2) can be undefined on
POSIX shm, replaced with ftruncate(2). IRMd check on existing lockfile
more portable.
FreeBSD 11.0 is preferred as it natively supports robust mutexes.
Full working LLC implementation pending.
|
|
|
|
|
| |
It was missing the FLOW_PENDING state. All states that are not
FLOW_ALLOCATED should return -1. This is now fixed.
|
|
|
|
|
|
|
|
| |
When a pending accept is shutdown on irmd exit, there are no more
threads running, but it should also change the state to NULL. This is
now correctly handled in the cleanup of the cancellation point.
Also fixed a busy wait with a condition variable.
|
|
|
|
|
|
|
|
|
| |
Flow allocation requests and registered api states revised so all
states are tracked with a condition variable. This is a more reliable
approach and improves stability of flow allocation.
Some other refactoring was also done, such as renaming port_map_entry
to irm_flow and hiding some internal structures of the registry.
|
|
|
|
|
| |
The destroy now only frees after all threads stopped using the object.
Also fixes a SEGV when trying to allocate a flow to a remote name.
|
| |
|
|
|
|
|
|
| |
This removes the parameter N-1 DIF from the enroll operation. IPCPs
should just allocate a flow to a DIF name and let the N-1 DIF resolve
it.
|